Ubuntu22.04 + Orin Arm Build Error: TENSORRT_NVPARSERS_LIBRARY variables are used in this project, but they are set to NOTFOUND.

[yanjingang]

Env:

Jetson AGX Orin 32G Board
Ubuntu 22.04
CUDA 12.6
CuDNN 9.3.0.75-1+cuda12.6
TensorRT 10.3.0.30-1+cuda12.5
Opencv 4.8.1

`$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Wed_Aug_14_10:14:07_PDT_2024
Cuda compilation tools, release 12.6, V12.6.68
Build cuda_12.6.r12.6/compiler.34714021_0

$ dpkg -l | grep -i cudnn
ii libcudnn9-cuda-12 9.3.0.75-1 arm64 cuDNN runtime libraries for CUDA 12.6
ii libcudnn9-dev-cuda-12 9.3.0.75-1 arm64 cuDNN development headers and symlinks for CUDA 12.6
ii libcudnn9-samples 9.3.0.75-1 all cuDNN samples
ii ros-humble-cudnn-cmake-module 0.0.1-3jammy.20240729.001922 arm64 Exports a CMake module to find cuDNN.

$ dpkg-query -W tensorrt
tensorrt 10.3.0.30-1+cuda12.5

$ opencv_version
4.8.1`

Build Error:

`$ colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release --packages-select=autoware_tensorrt_common
Starting >>> autoware_tensorrt_common
stderr: autoware_tensorrt_common
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
TENSORRT_NVPARSERS_LIBRARY
linked by target "autoware_tensorrt_common" in directory /home/work/autoware/src/universe/autoware.universe/perception/autoware_tensorrt_common
CMake Generate step failed. Build files cannot be regenerated correctly.

Failed <<< autoware_tensorrt_common [0.97s, exited with code 1]

Summary: 0 packages finished [3.82s]
1 package failed: autoware_tensorrt_common
1 package had stderr output: autoware_tensorrt_common`

Check:

libnvparsers not install, and orin arm env not have this lib
`$ dpkg -l | grep -i libnvinfer-dev
ii libnvinfer-dev 10.3.0.30-1+cuda12.5 arm64 TensorRT development libraries
ii python3-libnvinfer-dev 10.3.0.30-1+cuda12.5 arm64 Python 3 development package for TensorRT standard runtime

$ dpkg -l | grep -i libnvparsers-dev

$ sudo apt install libnvparsers-dev
E: Unable to locate package libnvparsers-dev`

Check Nvidia docs:
libnvparsers is removed when 9.0.1, last version of has libnvparsers is 8.6.1

tensorrt 9.0.1
https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt-901/release-notes/index.html

With removal of ICaffeParser and IUffParsers,
the libnvparsers library is removed.

tensorrt 8.6.1
https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt-861/install-guide/index.html#installing-debian
sudo apt install libnvparsers-dev

tensorrt 8.6.1 only support x86 ubuntu20, not support arm ubuntu22
xhttps://developer.nvidia.com/nvidia-tensorrt-8x-download

Question:

arm borad only can use ubuntu 20.04?
Tks!

[amadeuszsz]

Hello @yanjingang
Currently we don't support TensorRT 10.x. We are now working on CUDA upgrade, but it will take some time to apply necessary changes for Autoware. As a quick solution you can downgrade JetPack to 6.0 or try with Docker images,

arm borad only can use ubuntu 20.04?

JetPack 6.0 is based on Ubuntu 22.04 and contains CUDA 12.2, Tensorrt 8.6, cuDNN 8.9. It should work with Autoware.

[yanjingang]

Thanks a lot!I will try to downgrade to JatPack 6.0, Also wish you work well!Sent using the mobile mail appOn 2024/10/29 at 22:24, Amadeusz Szymko wrote: From: "Amadeusz Szymko" ***@***.***>Date: October 29, 2024To: "autowarefoundation/autoware" ***@***.***>Cc: "Mention" ***@***.***>,"yanjingang" ***@***.***>Subject: Re: [autowarefoundation/autoware] Ubuntu22.04 + Orin Arm Build Error: TENSORRT_NVPARSERS_LIBRARY variables are used in this project, but they are set to NOTFOUND. (Discussion #5379) Hello @yanjingang Currently we don't support TensorRT 10.x. We are now working on CUDA upgrade, but it will take some time to apply necessary changes for Autoware. As a quick solution you can downgrade JetPack to 6.0 or try with Docker images, arm borad only can use ubuntu 20.04? JetPack 6.0 is based on Ubuntu 22.04 and contains CUDA 12.2, Tensorrt 8.6, cuDNN 8.9. It should work with Autoware. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>